New paper on QuIP case selection strategies
20 May 2020 | Uncategorised|
We’re pleased to make available a pre-publication version of a new paper by James Copestake, forthcoming in the journal Development in Practice; Case and evidence selection for robust generalisation in impact evaluation
The paper focuses on sampling – or ‘case selection’, one of the areas researchers new to QuIP find most difficult to reconcile with the standard ways we are taught about sampling in relation to quantitative work. With the QuIP we often simultaneously pursue two strategies. First, we seek to identify as many causal mechanisms as possible, or achieve the highest possible degree of ‘saturation’, while recognising this is not something that can be quantified. Second we seek to add to the confidence we have in prior theory by applying the logic of ‘Bayesian uplifting’, while again rarely finding it useful to quantify this. With both strategies the key to maximising evidence generation within a budget is to make the best possible use of what we already know. James’ paper demonstrates how doing this transparently can rebuff the criticism that qualitative data is ‘merely anecdotal’ or worse still, biased by ‘cherry picking’.